DCGAN is initialized with random weights, so a random code plugged in the network would generate a very random image. Having said that, when you might imagine, the network has countless parameters that we will tweak, and the objective is to locate a placing of such parameters which makes samples g